Does anybody has any idea to find the inverse of a large matrix of size around 12000x12000 with float values?? If anybody has tried please help me.
thank you

too less information, too general problem. you have to know at least type of your matrix. different methods apply to different types. search google for keywords like sparse, band, diagonal, tridiagonal ect.

this may help also:
http://mathworld.wolfram.com/topics/MatrixTypes.html

too less information, too general problem. you have to know at least type of your matrix. different methods apply to different types. search google for keywords like sparse, band, diagonal, tridiagonal ect.

this may help also:
http://mathworld.wolfram.com/topics/MatrixTypes.html


Its a sparse matrix. when I'm using LU its taking more than 11hrs. please tell me the wayout to decrease the computational time to find the inverse of it. Even allocating dynamic memory still its taking that much time. please help me. Its urgent...... :confused:

There is chapter 2.7 of NR focusing to this type of matrices. Why do you try to use improper method as brute force?

thanks...
yah i read this chapter 2.7 but it doesnot exactly give the code for a sparse matrix inversion....it talks about storage of sparse matrix and for inversion it has given some references to which i'm not getting any access....
beauty123 if you already got any code for sparse matrix inversion please help me.......
My sparse matrix size is 12000x12000, and i don't know the pattern of the sparsity of the matrix.
its really urgent and i'm not getting any wayout
:confused:

regards

Thanks for this valuable thread!